SIDAR: Synthetic Image Dataset for Alignment & Restoration

Kwiatkowski, M.; Matern, S. and Hellwich, O. (2024). SIDAR: Synthetic Image Dataset for Alignment & Restoration. In Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 3: VISAPP, ISBN 978-989-758-679-8, ISSN 2184-4321, pages 175-189.

arXiv preprint: [https://arxiv.org/abs/2305.12036]

Dataset

• Fronto-Parallel View: [https://tubcloud.tu-berlin.de/s/55oX7tbXybCF9Be]

Prerequisites

• Blender 4.x
• Installation of Numpy inside Blender's API

Running the code

blender -b --python main.py -- <parameters>

Show help:
blender -b --python main.py -- -h

Example using default settings:

blender -b --python main.py -- --src <image-path> --dst <destination directory>